Appendix 1, proof of Proposition 2, p. 108 — under (d), r_0(t) ≤ f_0(t), so inventory is non-increasing
ProvedQFlexSC.ZeroInv.inventory_nonincreasingConsider a flex node running the Minimum Commitment (MC) policy (21)–(23). Assume the standing conditions and (), that all release quantities are nonnegative, , that the release schedules obey the output IR constraints (6), and that the input profile dominates the output profile componentwise (condition (d) of Proposition 2):
Let the run start from and the lot-for-lot schedule . Then for every
Together with the coverage property , this gives the first part of Proposition 2: inventory that starts at zero stays at zero.
Formalization Note. The paper derives from its formula (33), whose maximum over reaches periods before the start of a run; here the claim is stated directly for a run started at period . Two hypotheses are added and disclosed: (the paper's quantities are "measured in end-item equivalents"; the argument multiplies an inequality between flexibility ratios by ), and the initial state , (the paper states no start; with arbitrarily large, exceeds ). Parameter index is unused.
import Mathlib import Definitions.Def_QFlexSC_ZeroInv_Model
namespace QFlexSC.ZeroInv
theorem inventory_nonincreasing (P : QFParams) (f : ℕ → ℕ → ℝ) (I₀ : ℝ) (r₀ : ℕ → ℝ)
(hstd : ∀ q, 1 ≤ q → 0 ≤ P.αin q ∧ 0 ≤ P.αout q ∧ 0 ≤ P.ωin q ∧ P.ωin q ≤ 1 ∧
0 ≤ P.ωout q ∧ P.ωout q ≤ 1)
(hf : ∀ t j, 0 ≤ f t j)
(hIR : IROut P f)
(hI₀ : 0 ≤ I₀)
(hr₀ : r₀ = f 0)
(hd : ∀ j, 1 ≤ j → Acum P.αout j ≤ Acum P.αin j ∧ Ωcum P.ωout j ≤ Ωcum P.ωin j) :
∀ t : ℕ, 1 ≤ t →
sched P I₀ r₀ f t 0 ≤ f t 0 ∧ inv P I₀ r₀ f t ≤ inv P I₀ r₀ f (t - 1) := by sorry
end QFlexSC.ZeroInv
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.